Infrared radiation and energy evolution effects of coal rock under hydrodynamic coupling

Given the escalating issues of water intrusion in underground engineering, there is an urgent need for predictive warning systems regarding water-related disasters. This paper investigates the infrared radiation and energy evolution effects of coal-rock under hydromechanical coupling conditions through laboratory experiments, aiming to anticipate water outbursts resulting from the fracturing of coal-rock. Red sandstone specimens containing internal cavities with a diameter of 50 mm and a depth of 60 mm were subjected to water pressure of 0 MPa, 0.2 MPa, 0.4 MPa, and 0.6 MPa, applied internally via a hydraulic pump, while uniaxial loading was conducted in the vertical direction. This study analyzed the characteristics of Stress-Strain behavior, Energy evolution, Average Infrared Radiation Temperature (AIRT), and Variance of Successive Minus Infrared Image Temperature (VSMIT) during the rock failure and subsequent water outburst events. In addition, a novel indicator, the Pixel Standard Deviation of temperature (PSD), is introduced for enhanced assessment of these phenomena. The results indicate that water pressure significantly reduces the uniaxial compressive strength of the rock, with a decrease of 37.09 % observed at 0.6 MPa compared to 0 MPa. The ratio of elastic energy to dissipated energy (K_ED) evolves over time in a manner reminiscent of a peak shape, with distinct maxima that may serve as precursors to rock failure and water outbursts. The variations in the AIRT curve exhibit a high degree of consistency with those of the stress curve; specifically, as water pressure increases, the amplitude of AIRT fluctuations diminishes. At the moment of peak stress, the VSMIT experiences abrupt pulse changes, and the frequency of these VSMIT fluctuations is correlated with the occurrence of stress drop phenomena: an increase in the number of stress drop events corresponds to a greater number of VSMIT sudden changes. The PSD image stabilizes at approximately 70 % of the final failure time, at which point the proportions of stress at peak values under different water pressures are 67.76 % σ_p, 77.62 % σ_p, 81.54 % σ_p, and 53.69 % σ_p, respectively. Based on this, it can be utilized to predict the timing and morphology of coal and rock failure and water outbursts under varying water pressure conditions. Artificial Neural Network (ANN) model is effective for predicting rock failure and water outbursts; the predictions related to stress–strain behavior and K_ED from the neural network closely align with actual results, suggesting that temporal and infrared indicators can serve as effective inputs in forecasting sandstone failure and associated water outbursts.